9238
-40
I
1 3OOmp
35Omp (1'
lesa soluble Isomer
I
I
450mp
400 m p
500mp
b = mote soluble Isomer
Figure 1. Circular dichroism spectra of the diastereomers of [(/?6-CsHa)Fe(CO)(PPha)( C((S)-NHCHMePhIMe ]I[BF41.
trum whose positions clearly indicate that they arise from the chiral (h6-C5H,)Fe(CO)(PPh3) chromophore. It was noted above that secondary amines do not react with 1; presumably this is because of severe steric requirements near the iron center making it most reasonable t o assign the syn configuration to 8a and 8b. The metal center is stable t o racemization. The resolved complexes are unchanged on refluxing in acetone. This is the first example of the resolution of a chiral transition metal center in an organometallic compound which is stable to racemization and has the potential of retaining its optical activity during subsequent ligand transformations. In this connection, it should be pointed out that the first resolutions of transition metal compounds containing four different ligands by Brunner and coworkers'O were not amenable to further reactions without racemization. (10) H. Brunner, Angew. Chem., Inr. Ed. Engl., 10,249 (1971).
A. Davison,* D. L. Reger
NaBH, in ethanol gives two products in approximately equal amounts. One of these, 4, is that expected from hydride attack at the carboxonium carbon; the other (6). These is the a-ethyl (h5-C5H5)Fe(CO)(PPh3)(CzH5) two complexes can be separated easily, 6 being considerably less soluble in hexane.' The reaction of 1 with NaBD4 in EtOH gives (h5-C5H5)Fe(CO)(PPh3){ CD(OEt)CH3} (4a) and (h5-C5H5)Fe(CO)(PPh3)(CD2CHs) (6a) which shows that all of the added hydrogens in the formation of 4 and 6 arise from the borohydride. The formation of 6 is difficult to explain; it has been found that 4 does not react with BzHe or BzHB-NaBH4 mixtures in EtOH; however, [Ph(COEt)NEtz][BFa] reactss with NaBH4 to give PhCHzNEtz which again supports the contention that 1 behaves as a carboxonium salt. We have synthesized 6 by an independent route: (h5-C5Hj)Fe(CO)(PPh3)Iwas treated with AgBF4 and ethylene to give [(h5-C5H5)Fe(CO)(PPh3)(hzC2H4)]BF, (7); reduction of the latter with NaBH, yields 6. Primary amines and ammonia, but not secondary amines, react with 1 to give carbimonium salts; e.g., (S)-(-)-a-phenylethylamine ( [ a I z 5 ~- 39.6') gives a mixture of the diastereomers [(h5-C5H6)Fe(CO)(PPh3){ C((S)-NHCHMePh)Me)]BF4 (sa and 8b). Resolution of the RS and SS isomers was achieved by fractionation from ethanol. As with 4, the separation can be conveniently followed by nmr in the methyl region : nmr (6 in acetone-&) less soluble isomer (mp 195' dec) 1.61 (3, doublet, J = 6.8 Hz, CHMe), 2.85 (3, singlet, CMe); more soluble isomer (mp 186" dec) 1.44 (3, doublet, J = 6.8 Hz, CHMe), 2.92 (3, singlet, CMe). Conclusive proof that the fractionation was in fact a resolution and not the separation of syn-anti isomers about the CN multiple bondg comes from the circular dichroism spectra shown in Figure 1. Three Cotton effects are found in the 500-300-mp region of the spec(7) The reduction of 1 with NaBH4 was previously reported.' However, these authors failed to notice the presence of the markedly less soluble 6 which is not eluted by chromatography in hexane; further, 4 and 6 have identical carbonyl stretches in the ir spectrum, v (CO) 1912 cm-l (hexane), such that the presence of 6 in the crude reaction mixture is not easily ascertained. (8) R. F. Borch, TetrahedronLert., 61 (1968). (9) This type of isomer has been found in Cr(C0)a (C(NHMe)Me) : E. 0. Fischer, B. Heckl, K. H. Dotz, and J . Muller, J . Organometa/. Chem., 16, P29 (1969).
Journal of the American Chemical Society
/
94:26
/
Department of Chemistry, Massachusetts Institute of Technology Cambridge, Massachusetts 02I39 Recehed August 23, I972
Chemiluminescence from the Reaction of the Hydrated Electron with Tris(bipyrid yl)ruthenium(III) Sir : The hydrated electron, eaQ-, is a rapid, powerful, and simple reducing agent which on occasion has been postulated to give an excited-state product. 2 , 3 Undeniable evidence that such a product is formed would be that emission characteristic of a known product excited state is seen. We report here apparently the first observation of such emission along with an estimate of the chemiluminescent yield, &l. The reaction is most simply represented by eq la,b, where bipy denotes bipyRu(bipy)P
+
/ \
+ hv
[ R ~ ( b i p y ) , ~ ++ ] * Ru(bipy)a2+
(la) transition
[
state
1* +R ~ ( b i p y ) , ~ +( l b )
ridyl. The strong room-temperature photoluminescence of [ R ~ ( b i p y ) ~ ~ at + ] *ca. 630 nm has been assigned as phosphorescence from a triplet charge transfer, 3CT, This same excited state can function as a sensitizer, either by excitation energy transferB or as an excited-state reducing agent.9 The state may also be produced chemically, by reduction of R ~ ( b i p y3+) ~by hydroxide ionlo and by hydrazine," and derived from electrogenerated species in acetonitrile solutions of [Ru(1) E. J. Hart and M. Anbar, "The Hydrated Electron," Wiley, New York,N. Y., 1970. (2) W. L. Waltz and R. G. Pearson, J . Phys. Chem., 73, 1941 (1969). (3) M. 2. Hoffman and M. Simic, J . Amer. Chem. Soc., 94, 1757 (1972). (4) D. M. Klassen and G. A. Crosby, J . Chem. Phys., 48, 1853 (1968). (5) J. N. Demas and G. A. Crosby, J . Mol. Speclrosc., 26,72 (1968). (6) F. E. Lytle and D. M. Hercules, J . Amer. Chem. Soc., 91, 253 ( 1969). (7) J. N. Demas and G. A. Crosby, ibid., 93,2841 (1971). (8) J. N. Demas and A. W. Adamson, ibid., 93, 1800 (1971). (9) H. Gafney and A. W. Adamson, unpublished work. (10) D. M. Hercules and F. E. Lytle, J . Amer. Chem. Soc., 88, 4745 (1966). (1 1) (a) F. E. Lytle and D . M. Hercules, Photochem. and PhotobioL, 13, 123 (1971); (b) D . M. Hercules, Accounts Chem. Res., 2,301 (1969).
December 27, 1972
9239
(bipy),]Clz. l 2 The present investigation was undertaken to see if chemiluminescence could be produced by means of a single-step reaction between well-defined reactants. Sample handling techniques were essentially as described elsewhere. l, Concentrated Ru(bipy)a 3+ solutions were chemically generated by PbO, or Clz oxidation of Ru(bipy),Clz14in aqueous 0.1 N H2S04,filtered, saturated with He, and injected into the matrix yielding a final [Ru] of (0.2-1.0) X 10-3 M . The matrix consisted of triply distilled water, 0.2 M in tert-butyl alcohol for rapid .OH scavenging (tI/* 11 nsec), and 1 m M HOAc-2mM NaOAc for buffering. The pH of the irradiated solutions thus prepared was approximately 4.5, therefore limiting the reaction, Ru(bipy)s3+ * H (formed from eaqH+, tl/, 1 psec), to a minor role in r e d u ~ t i 0 n . l ~eaq- was effectively the only radiation induced species involved and was observed to disappear 10-60 nsec from the irradiated solutions with tl/, (corresponding to k[,,,-+Ru(bipy)sai] 'u 6 X 1O1O M-' sec-I). The hydrated electrons were produced in the He or NzO saturated solutions by 4-40 nsec single pulses of 15 MeV electrons from the Argonne National Laboratory's A R C 0 Linac. Emission from the pulsed samples was monitored by an HTV R213 photomultiplier and Bausch and Lomb high-intensity monochromator combination. The spectrum, corrected for photomultiplier-monochromator response, is shown in Figure 1. The radiative lifetime, T, for the ea,R ~ ( b i p y ) ~produced ~+ emitter was determined at all experimentally monitored wavelengths and consistently found to be 0.58 psec at 25". For comparison the emission spectrum of R~(bipy)3~+ (excited at 453 nm) is also given.l6 7 , obtained by photoexcitation in aqueous 0.2 M tert-butyl alcohol with a pulsed Nz laser, is 0.63 psec. That the emission arises from the reaction of the eaq- with the complex is demonstrated by the greater than sixfold reduction of emission intensity upon NzO 4.9 nsec. saturation of the matrix tI/,(NzO+e,,-) eerenkov stimulation of emission from Ru(bipy), 2+ (produced prior to the Linac pulse by thermal reduction) can be relegated to a minor role, since emission intensity from a pulsed 2 X M Ru(bipy),,+ solution was only 0.1 that observed on identically pulsing a 2 X M R ~ ( b i p y ) ~sample. ~+ The yield of Ru(bipy)s2+, determined from its absorption at 453 nm, is (1.0 f 0. l)/eaq-; the dose absorbed by the solution was evaluated by "Super" Fricke dosimetry. 4cl has been estimated as 0.015 f 0.004 by compar-
-
+
+
-
+
-
(12) N. E. Toke1 and A. J. Bard, J . Amer. Chem. SOC.,94, 2862 (1972). (13) J. K. Thomas, S. Gordon, and E. J. Hart, J . Phys. Chem., 68, 1524 (1964). (14) Prepared by the method of J. E. Ferguson and G. M. Harris J . Chem. Soc. A, 1293 (1966). (15) In separate experiments conducted at lower pH, we have determined the rate constants for the reduction of Ru(bipy)aS+ by H ' atoms and (CH*)(CH&C(OH) radicals to be 1.7 X 109 and 1.3 X 108 M-l sec-1, respectively. While the reduction by H also produces chemiluminescence (of as yet undetermined quantum yield), both of these reactions are too slow to account for the chemiluminescence described here and attributed to the hydrated electron. (16) Photoluminescence was produced by W lamp output passed through a Bausch & Lomb monochromator and, at go", chopped and analyzed by a second monochromator and detected by an RCA C3 1025C photomultiplier, whose output was monitored by a Brower Lock-In voltmeter synchronized with the chopper. (17) H . Fricke and E. J . Hart, Radiar. Dosimetry, 2 , 167 (1966).
-20
18
16
14
12
F (cm-' x
+
Figure 1. Chemiluminescence spectrum of Ru(bipy)s3+ eaq(0-4 and )photoluminescence spectrum of R u ( b i p y ) P (.-e).
ing the detector response at 620 nm from a photoexcited R ~ ( b i p y ) ~ , +solution of known photoluminescence yield, 4p1,with the response obtained on reduction of R ~ ( b i p y3+) ~by a known quantity of hydrated electrons1s (the optical and detector systems being unchanged). This value of &I, which is of the same order of magnitude as $1~1, is relatively high for an in vitro chemiluminescent system. 2o We believe that these results do establish that reaction 1 indeed occurs and that the produced excited state is essentially the same 3CT state observed in emission studies. We also find that the temperature variation of the emission yield (8-35') corresponds to an activation energy of 1.5 kcal mol-'; this compares with the value of 2.5 kcal mol-' for the radiative decay following photoexcitation of Ru(bipy),,+ in water.6 Additional quantitative measurements of +cl and of the reaction kinetics are now in progress. The production of the chemiluminescent state [Ru(bipy),Z+]* (with an excitation energy of ca. 2.0 eV) is consistent with the energetics of the reduction of Ru( b i ~ y ) , ~by+ the powerful reducing agent eaq--,as well as by electrogenerated R ~ ( b i p y ) ~ + ,but not by weaker chemical one-electron reducing agents such as TiC13, which indeed failed to yield chemiluminescence, la or C r ( b i p ~ ) ~ ~which +, produced only very weak chemiluminescence. 21 In this context the strong chemiluminescence reported l 1 to accompany the reduction of Ru(bipy)33+ by the relatively weak reducing agents OH- and N,HJ is somewhat surprising and is presumarly associated with steps involving high-energy intermediates (e.g., HOz. and N2Hz) in the complex oxidations of these reductants to 0, and Nz, respectively. Noteworthy in this respect is our preliminary observation that quite strong chemiluminescence ap(18) A40-nsecpulseof derenkovlight excited the aqueousRu(bipy)~2+ a t 25". The intensity of absorbed light was measured by ferrioxalate actinometry.19 The value of &l used was 0.02 which was determined by comparison of corrected spectra for Ru(bipy):%+and quinine ~ steady illumination at 365 nm. bisulfate in 1.0 N H z S O under (19) C. G . Hatchard and C. A. Parker, Proc. Roy. Soc., Ser. A , 235, 518 (1956). (20) K. D. Gundermann, Angew. Chem.,Znt. Ed. Engl., 4,566 (1965). (21) 3. Halpern and P. F. Phelan, unpublished results.
Communications to the Editor
9240 pears on mixing aqueous NaBH4 with an acid solution of R ~ ( b i p y ) ~ ;3we + also find luminescence on reduction of R ~ i ( b i p y ) ~by~ aqueous + H . atoms. Acknowledgment. This work was performed under the auspices of the U. S . Atomic Energy Commission, with support through Grant GP-26600 to the University of Chicago from the National Science Foundation and through Contract DA-ARO-D-31-124-G87 between the U. S . Army Research Office (Durham) and the University of Southern California. We thank the Argonne Linac staff for experimental assistance. John E. Martin,* Edwin J. Hart Chemistry Division, Argonne National Laboratory Argonne, Illinois 60439
Arthur W. Adamson, Harry Gafney Department of Chemistry, Unicersity of Southern California Los Angeles, California 90007
Jack Halpern Department of Chemistry, University of Chicago Chicago, Illinois 60637 Receiced May 16, 1972
Evidence for Stepwise Solvation of Strontium Carbanion Salts in Ethereal Solvents Sir : The structure of ion pairs of alkali salts of carbanions and radical anions in aprotic solvents has been extensively studied in the last few years. The existence of contact or tight ion pairs and solvent-separated or loose ion pairs in these systems is now well documented.1-3 Only recently has information on the structure of alkaline earth carbanions become available. 4 4 In a previous communication we reported on the structure of fluorenylbarium and its crown ether complex in a few aprotic solvents such as tetrahydrofuran (THF) and dimethoxyethane (DME).4 It was shown that this salt in T H F over the temperature range 25 to -70" exists essentially as a tight ion pair. We now want to report evidence for a distinct stepwise solvation of the fluorenylstrontium ion pair in T H F in the temperature range 20 to -100". This process involves the formation of an asymmetrically solvated ion pair in which one of the fluorenyl moieties is separated from the strontium ion followed by the separation of the second fluorenyl unit t o give a symmetrically solvated ion pair. F1-M2+F1-
__
Fl-//MZ+Fl- f, Fl-~\M2+llFl-
The strontium salt was prepared in T H F over a strontium mirror in the presence of a stoichiometric quantity of 1,l -diphenylethylene in an all-glass apmm) conditions. paratus under high vacuum ( The visible spectrum of the recrystallized salt has max(1) T. E. Hogen-Esch and J. Smid, J . Amer. Cltem. Soc., 87, 669 (1965); 88, 307, 318 (1966). (2) R. C. Roberts and M. Szwarc, ibid., 87, 5543 (1965); R. V. Slates and M. Szwarc, ibid., 89, 6043 (1967). (3) E. de Boer, Recl. Truu. Chim. Puys-Bus, 84, 609 (1965); N. Hirota, R. Carraway, and W. Schook, J . Amer. Chem. Soc., 90, 3611 (1968). (4) T. E. Hogen-Esch and J. Smid, ibid., 91,4580 (1969). (5) J. C. Favier, M. Fontanille, and P. Sigwalt, Bull.Soc. Chim. Fr., 2, 526 (1971); J. P. Pascault and J. Gole, J . Chim. Phys., 68, 442, 449 (1971); W. T. Ford, J . Organometal. Chem., 32,27 (1971).
Journal of the American Chemical Society J 94:26
/
ima at 343, 435, and 458 nm and indicates a contact ion pair (Figure 1). On lowering the temperature the small shoulder at 372 nm, identifiable with separated ion pairs, increases rapidly in intensity. At -22" the two bands at 343 and 372 nm are approximately equal in intensity but at lower temperatures a large absorption band at 343 nm persists (Figure 1).6 This behavior contrasts sharply with that of the alkali salts such as fluorenylsodium in this solvent where the transition from contact ion pairs to separated ion pairs is essentially complete at -70" and the contact ion pair absorption has disappeared. The latter process is relatively straightforward R
NafFl-
AH = -7.6 kcal
Na+llFl-
and the log R us. 1/T plot is essentially straight over the entire temperature range ( R = [separated ion pairs]/ [contact ion pairs]). In the present case, however, the log R us. 1/T plot is strongly curved (Figure 2).7 Apparently the process is strongly exothermic ( A H 'v - 12 kcal), until approximately 50% of the contact ion pairs have been converted into separated ion pairs. The apparent enthalpy of the solvation process then decreases to a much lower value (AH 'v -0.5 kcal) and increases again at lower temperatures. These observations suggest that the process is not represented by a simple one-step solvation process K8
Fl-Sr2+Fl-
Fl-//Sr2+//F1-
but possibly by a two-step process in which the fluorenyl ions become successively separated from the Sr 2+ ion. The first step is expected to be highly exothermic, the second step much less so. Kz
KI
F1-Sr2+F1-
F1-(/Sr2+F1-
Fl-!lSr2+llFl-
From this scheme it can be seen that the ratio of [separated ion pairs] to [contact ion pairs] is given as shown in eq 1. R = (1 2&)/(1 2/K1) (1 1 Taking the logarithm and differentiating with respect to 1/T gives the apparent enthalpy of the process, contact ion pairs e separated ion pairs (eq 2). From
+
AH
(2/(2
+ Ki))AHi +
+
+ 2Kz))AHz
(2)
(2) it follows that at high temperature, where KI and K&< 1 , A H E A H l . (6) Complete reversibility of the spectral phenomena and very low conductance of the solution assure that we are again dealing with ionpair equilibria. (7) The ratio R can as before (see ref 1) be evaluated using spectra of carbanion solutions existing as contact ion pairs or as separated ion pairs over the entire temperature range, I n this case data on fluorenylbarium in T H F and fluorenyllithium in DME were used to describe the behavior of contact and separated ion pairs. R was calculated on the assumption that the spectra of the two "types" of separated ion pairs, e.g., of 11F1- in FI-SrI /F1- and in FI-lISr/ IFI-, are identical. Extinction coefficients (e) of contact and separated ion pairs are taken as 10,100 and 14,000, respectively, at 25". The change of extinction was detercoefficient with temperature expressed as - d(eTo/ezoo)/dT mined as 4.0 X 10-3 and 6.0 X 10-3 for contact and separated ion pairs, respectively. The contribution of separated ion pairs at 343 nm defined as ODardODmax312 and the contribution of the contact ion pairs at 373 nm (oDa,z/oD,., 343) were determined as a function of temperature. At -76", for instance, these values were 0.26 and 0.27, respectively. Using these data one calculates that the increase of OD of the contact ion pair at 343 nm on lowering the temperature is only apparent. For instance, the corrected absorption of the contact ion pair at -76" is only about 0.60.
December 27, 1972